JOURNAL OF BACTERIOLOGY, Apr. 1974, p. 59-69Copyright 0 1974 American Society for Microbiology

Vol. 118, No. 1Printed in U.S.A.

Microtubule Assembly and Function in Chlamydomonas:Inhibition of Growth and Flagellar Regeneration by

Antitubulins and Other Drugs and Isolation of ResistantMutants

MARTIN FLAVIN AND CLARENCE SLAUGHTERLaboratory of Biochemistry, National Heart and Lung Institute, National Institutes of Health, Bethesda,

Maryland 20014

Received for publication 26 October 1973

The distribution of microtubules in Chlamydomonas reinhardtii suggests thatthey are involved in mitosis, cell and nuclear cleavage, and generation of flagella.Vinblastine, colchicine, and podophyllotoxin bind to the protein building blockof microtubules (tubulin) and prevent normal assembly. Mutants resistant tothese "antitubulin" drugs are candidates to have alterations in tubulin primarystructure. We report the ability to inhibit growth, and flagellar regeneration afteramputation, of: vinblastine, several colchicine derivatives, two water-solublederivatives of podophyllotoxin (succinylpodophyllotoxin and epipodophyllotoxinthiuronium bromide), and other substances which may interfere with flagellarassembly or motility (isopropyl N-phenyl carbamate, 2-methoxy-5-nitrotropone,chloral hydrate, caffeine, and nickel acetate). The ability of each drug to inhibitbinding of labeled colchicine or podophyllotoxin to mammalian brain tubulinwas also determined. The results suggest that only in the cases of colchicine,colcemide, and epipodophyllotoxin thiruonium bromide was the toxicity toChlamydomonas mediated by inhibition of tubulin assembly. The requirementfor high concentrations of colchicine may be due to permeability barriers, sincecolchicine toxicity was potentiated by deoxycholate. Mutants resistant toantitubulins were isolated after treatment with methyl methanesulfonate. Theresults with vinblastine were equivocal. Of three mutants resistant to inhibitionof growth and flagellar regeneration by colchicine, one was also cross-resistant toepipodophyllotoxin thiuronium bromide.

Cilia and simple eukaryotic flagella sharemany common biochemical and ultrastructuralproperties, and advances in the past decade (forreviews see references 2, 13, 20, 24) have nowmade feasible studies of their assembly andmotility, and the mechanisms for regulatingassembly and motility. Chlamydomonas rein-hardtii, a unicellular biflagellate alga, is espe-cially suited for genetic investigation of theseproblems (10). Its flagella contain the charac-teristic set of nine outer doublet microtubulesarrayed about a central pair of singlet tubules.The microtubules are polymers of two or moreclosely related proteins (tubulins), of molecularweight about 55,000 (26). Suitable low-tempera-ture extraction dissociates cytoplasmic mi-crotubules into dimeric (probably heterodi-mers) precursors, and a precursor pool is alsopresent in Chlamydomonas (15). Colchicineand some other antimitotic agents bind to the

dimer, preventing polymerization and provid-ing a convenient assay method.The utilization of adenosine 5'-triphosphate

(ATP) for flagellar motility is believed to bemediated by a heterogeneous ATPase, dynein,which is located in sidearms of the outer doublettubules (5). Recently, we have also identified inChlamydomonas flagella: (i) a low molecularweight, calcium-specific ATPase which is ofinterest, although a role in motility is notestablished, because calcium ions may partici-pate in the regulation of both microtubule as-sembly and flagellar motility (21); and (ii) anucleoside diphosphokinase activity which mayfunction in microtubule assembly (T. Wata-nabe and M. Flavin, Abstr. 719, Proceedings ofthe 13th annual meeting of the American Soci-ety for Cell Biology, November 1973).

Previous genetic approaches to flagellar as-sembly and function have been directed to the

59

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

FLAVIN AND SLAUGHTER

isolation of paralyzed mutants. At least 18genes have been identified, mutation in whichleads to paralysis or abnormal motility inChlamydomonas (12). These mutants haveplayed little part in recent advances, probablybecause their flagellar ultrastructure and chem-istry has not been systematically studied. It alsoappears that the phenotype, judged by lightmicroscopy, is sometimes nearly as variablewithin a population of one mutant as betweenmutants from different genes (12).We have begun to isolate and study mutants

which are resistant to drugs that bind to tubulinor affect microtubule assembly. We hope also tostudy mutants resistant to paralyzing drugs ordrugs which specifically inhibit dynein or theother enzymes mentioned above. Drugs of thelatter type, comparable to ouabain and oli-gomycin for Na/K and mitochondrial ATPase's,respectively, are not yet known, and if availablewould additionally be useful in defining thefunctions of flagellar enzymes.We first examined the ability of a number of

drugs to inhibit flagellar regeneration. Rosen-baum et al. (15) found that colchicine wasinhibitory, but at 1,000 times higher concentra-tion than required for binding to cytoplasmictubulin. This could be due to permeabilitybarriers, about which little is known inChlamydomonas. At the same time it has beendifficult to demonstrate colchicine binding toflagellar tubulin, probably because the outerdoublet microtubules do not readily dissociateinto dimers (25). Since it is not certain thatChlamydomonas flagellar microtubule precur-sors have the same drug affinity spectrum asthat established for cytoplasmic tubulin, wehave tried to compare binding to brain tubulinwith inhibition of flagellar regeneration forvarious colchicine and other derivatives, andthus confirm whether the effects ofChlamydomonas are mediated by interactionswith tubulin.

Microtubules are present not only in theflagella of Chlamydomonas but also in themitotic spindle, and in the cytoplasm wherethey participate in nuclear and cell division (8).Thus, agents preventing microtubule assemblyshould inhibit growth as well as flagellar regen-eration. Drugs of particular interest are: isopro-pyl N-phenyl carbamate, a herbicide whichallows microtubules to assemble but in a disor-ganized pattem (7); vinblastine, which binds totubulin dimer and induces abnormal aggrega-tion (14); and colchicine and podophyllotoxin,which bind to tubulin and inhibit assembly (13,20, 24). The latter three antimitotic agents can

specifically be called antitubulins, and presum-ably bind to or alter protein-protein interactionsites. Mutants resistant to such drugs would becandidates to have alterations in tubulin pri-mary structure. The antitubulins do not inducedepolymerization of flagellar microtubules;therefore, in evaluating the mechanism of drugaction it was important to distinguish betweenthe detachment of preformed flagella and inhi-bition of regeneration. (A preliminary report ofthis work appeared in Fed. Proc. 32:642, 1973.)

MATERIALS AND METHODSCulture conditions. A high salt minimal inorganic

medium, M, was used for all liquid cultures (19). Forsolid media this was supplemented with 1.5 to 2% agarand 0.2% sodium acetate, MA, or acetate plus 0.4%yeast extract, MAY. Solid media were color codedwith McCormick food colors; 0.1 ml/liter did notaffect growth or viability. Quantitative studies ofgrowth inhibition by drugs and other substances weredone by using 1 ml of solid medium dispensed into16-mm diameter wells in Linbro sterile plastic trays(Bellco Glass, Inc.). A 0.5-ml amount of variousconcentrations of Millipore filter-sterilized drug wasadded first, followed by 0.5 ml of double strength MAcontaining 3% agar; the latter was kept in a 60 Cwater bath for 30 min after autoclaving, and dis-pensed from a pipette briefly warmed in a flame. Thetrays were prepared the day before use. Petri dishesfor isolation and cloning of mutants were prepared byadding filter-sterilized drugs to MA or MAY mediumwhich had been cooled to 45 C after autoclaving. Insome cases, immediately after inoculating the solidi-fied medium, 4 ml of drug-containing soft agar (0.6%),cooled to 45 C, was added and spread by gentleswirling. For liquid cultures, 150 ml of M was auto-claved in 300-ml Erlenmeyer flasks closed with plasticfoam plugs (Arthur Thomas) through which a cotton-plugged Pasteur pipette had been inserted. A mixtureof 3% CO2 in air, prepared with a Matheson model665-1 gas proportioner equipped with 602 and 610flowmeters, was bubbled through each flask at a rateof about 100 ml/min.

Chlamydomonas reinhardtii wild-type 137c (mt+), obtained from R. P. Levene, was maintained onslants of MA and MAY in screw-top tubes. Bothparental and mutant strains were transferred every 3weeks and stored in the dark at +5 C. Cells werecultured at room temperature (24 to 28 C) andilluminated at 400 footcandles with cool-whitefluorescent light. Cultures on solid medium and liquidcultures to be used for mutagenesis were grown incontinuous light. Cells were harvested for mutagene-sis at a cell density of 106/ml. Liquid cultures for allother experiments were synchronized by a cycle of 14h of light and 10 h of dark (15). In practice, a liquidculture was diluted to 2 x 105 cells/ml every day, andcould then be harvested at 8 x 105 cells/ml between h2 and 4 of the light cycle the following day. A heavyinoculum, 0.05 ml = 40,000 cells, was spread over themedium in the plastic tray wells. When necessary,

60 J. BACTERIOL.

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 118,1974 CHLAMYDOMONAS MUTANTS RESISTANT TO ANTITUBULINS

cells were diluted in medium M, or concentrated bycentrifuging for 5 min at 1,000 x g at 25 C andresuspending in M.

Measurement of flageliar regeneration and de-tachment, and motility. To determine whether drugscaused flagellar detachment, an aqueous solution ofthe drug was mixed in a test tube with an equalvolume of cells at a density of 4 x 106/ml, and thetube was agitated for 2 h at 25 C under constantillumination, by gentle rocking. At intervals, portionswere fixed with glutaraldehyde (15), cell density was

determined with a Levy-Hausser counting chamber,and 100 cells were scored with a phase-contrastmicroscope for the presence of flagella or visibleflagellar abnormalities. To determine whether drugsinhibited motility (in the absence of flagellar detach-ment), portions were examined in the counting cham-ber with and without glutaraldehyde fixation. Sinceonly the percentage of cells which appeared totallyimmobile was recorded, the assay was not very

precise.To study flagellar regeneration, flagella were re-

moved by shearing with a Virtis-23 homogenizer (15).With 3 ml of cell suspension in the smallest flutedglass cup, 6 min at 25 C at the highest rheostat settingwas needed for complete amputation. Flagella were

then allowed to regenerate in gently rocking testtubes, as above. At 10 and 60 min (when regenerationwas essentially complete) portions were fixed withglutaraldehyde and the lengths of flagella from 50cells were measured under a phase microscope with anocular micrometer. Drugs to be tested for ability toinhibit regeneration were added immediately beforeor after amputation, as indicated in the text.

Drug-resistant mutants. In preliminary experi-ments several mutagens were studied by using theincidence of streptomycin resistance (9) to compare

their effectiveness. The conditions were similar tothose previously described for: ultraviolet light (10),N-methyl-N'-nitro-N-nitrosoguanidine (9), and ethyland methyl methanesulfonate (11). The latter was

then chosen and used as follows. A 400-ml amount of aculture grown to 6 x 105 to 10 x 10' cells/ml incontinuous light was centrifuged in sterile tubes for 6min at 1,000 x g at 25 C, and the pellet was washedonce with 0.02 M potassium phosphate, pH 6.0, andsuspended in 10 ml of this buffer. A 0.2 M solution ofmethyl methanesulfonate was prepared in the same

buffer and sterilized by filtration through a 0.2-aimpore size membrane. A total of 3 x 106 cells was

suspended in 25 ml of phosphate buffer with methylmethanesulfonate at a concentration of 0.016 M, andthe flask was kept on a rotary shaker, under 250footcandles of light, for 2 h at 25 C.

Cells were collected from 4-ml portions on 0.8 Mm,47-mm diameter Millipore filters (AAWP) andwashed twice with 10 ml of phosphate and once withmedium M to remove the mutagen. With thisamount of cells the washing took about 1 h, usingsuction from either a water pump or high vacuum oilpump. Before the cells were completely dry on thefilter they were suspended in 2 ml of medium M byagitation with a sterile Pasteur pipette. Portions were

plated on drug-containing medium (see below) atonce and, to allow "gene expression," also after: (i)diluting a portion with medium M to 107 cells/ml androcking the suspension in dim light (200 footcandles)at 25 C for 20 h, and (ii) inoculating 150 ml of me-dium M with 1 ml of the suspension and culturingin 400 footcandles continuous light for 2 days. Therewas no measurable cell increase in (i), due to the highinitial density, and about 20-fold increase in (ii);mutant colonies derived from (ii) could not be as-sumed to be of separate origin. Cells from (ii) wereconcentrated to 2 x 107/ml, by centrifugation, beforeplating.

Drug-containing plates were prepared the day be-fore use by adding filter-sterilized drug solutions toMA or MAY agar medium, cooled to 45 C, afterautoclaving, to give the following concentrations:streptomycin, 50 Ag/ml; colchicine, 10 mM; vinblas-tine, 0.5 mM; epipodophyllotoxin thiuronium bro-mide (EPT), 1.0 mM; isopropyl N-phenyl carbamate,0.8 mM; and nickel acetate, 0.5 mM. These concen-trations are half those of saturated aqueous solutionsin the case of EPT and the carbamate derivative. Thepetri plates were inoculated with 0.5 x 101 to 10 x106 total cells, depending on the estimated viability.Viability was determined at each step of the proce-dure by plating serial dilutions, in M, on MA me-dium; colonies could be counted after 7 days or 4 dayswith a dissecting microscope. A heavy inoculumcould be spread more uniformly by adding 4 ml of0.6% agar MA medium, cooled to 45 C; this some-times seemed also to increase the recovery of mutants.Mutant colonies picked for further study were

cloned by first transferring to MAY slants and thensubculturing in 150 ml of minimal M liquid medium,from which a few hundred cells were spread on platesofMA containing either 5 mM colchicine or 0.15 mMvinblastine. Single colonies were picked for furtherstudy.

Materials. Methyl and ethyl methanesulfonatewere obtained from Eastman, N-methyl-N'-nitro-N-nitrosoguanidine and caffeine from Aldrich, strep-tomycin and sodium deoxycholate from Calbiochem,chloral hydrate from Fisher, -vinblastine from Lilly,griseofulvin from Ayerst, and Brij 58 from AtlasChemical.

Colchicine was obtained from Calbiochem, andderivatives (4) as follows. Colcemide (Ciba) has anN-methyl group in place of N-acetyl. Desacetylcolchi-cine d-tartrate (Microbiological Associates, Bethesda,Md., SKF no. 250-F), with a free amino group, is alsocalled "trimethyl colchicinic acid methyl ether"; in apreliminary communication (Fed. Proc. 32:642, 1973)it was erroneously called trimethyl colchicinic acid.Desacetylthiocolchicine (M. Zweig and C. Chignell,NHLI, NIH) is the preceding compound with sulfurreplacing oxygen in the ring C methoxyl group.Colchicosamide (K and K Laboratories) is a glucosideof a colchicine derivative with an amino group replac-ing the methoxyl of ring C. 2-Methoxy-5-nitrotropone(H. Scheour, Calbiochem) is a possible analogue ofthe C ring of colchicine. 2-Hydroxy-5-nitrotroponewas prepared by heating a saturated aqueous solution

61

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

FLAVIN AND SLAUGHTER

(pH 4) of the preceding compound (molar absorbance11,000 at 350 nm) at 100 C until the spectral changewas complete (molar absorbance 20,000 at 440 nm).

Podophyllotoxin (Aldrich) was estimated to be 90%pure from its optical rotatory dispersion spectrum (6),and from thin-layer chromatography (M. Flavin,unpublished data). Succinylpodophyllotoxin, a wa-

ter-soluble derivative in which the benzylic hydroxylis esterified with succinic acid, was prepared andcrystallized as the isopropylammonium salt, as previ-ously described (17). Another water-soluble deriva-tive, EPT, melting point 149 to 151 C (obtained fromMicrobiological Associates, Bethesda, Md.) was pre-

sumably from the original preparation (17). The "epi"configuration was inferred (17) from the method ofpreparation, but has not been proven.

RESULTS

Permeability to antitubulins and inhibitionof flagellar regeneration. After amputation offlagella from cells isolated from a synchronousculture at the same stage of the cell cycle (15)flagella reproducibly regenerated to a third oftheir original length in 10 min, and to theiroriginal length, 12 ,im, in 60 min (Table 1).When added before amputation 1.5 mM colchi-cine completely inhibited regeneration, butwhen added immediately after, regenerationproceeded normally for the first 10 min and was

then arrested. When added after a 10-minregeneration period even much higher concen-

trations did not inhibit further regeneration(Table 1). A comparable lag in the onset ofinhibition was not observed (Table 1) withcolcemide or vinblastine. Although there was no

loss of viability under the above conditions, the

shearing may have done something more thanamputate flagella. There was a change in pH ofthe cell suspension, and there sometimesseemed to be two cell populations, one whichregenerated flagella relatively well and one

which did not regenerate at all (Tables 1 and 2).Table 2 shows that colchicine was 30 times

more effective in inhibiting flagellar regenera-

tion in the presence of a concentration ofdeoxycholate which by itself had no effect, i.e.,0.05 mM gave the same inhibition with deoxy-cholate as 1.5 mM without. This also suggeststhat a permeability barrier may be responsiblefor the requirement for relatively high concen-

trations of colchicine to inhibit flagellar regen-eration. A neutral detergent, Brij 58, partiallyinhibited flagellar regeneration at a concentra-tion of 1%, but did not potentiate the inhibitionby colchicine; i.e., the inhibition by both to-gether was equal to the sum of the inhibitionsby each separately. Flagellar regeneration was

also quite sensitive to organic solvents; 2%ethanol inhibited partially. This handicappedattempts to study the effects of water-insolubledrugs.

Ability of antitubulins and other sub-stances to inhibit growth, flagellar regener-ation or motility, or induce flagellar de-tachment. Table 3 summarizes the effects ofa number of substances, implicated in mi-crotubule assembly or flagellar function, on

Chlamydomonas in relation to four differentparameters. Initially we had hoped to find thatsome drugs would affect only flagellar regenera-

tion, but none were found that were not equally

TABLE 1. Time course of flagellar regeneration in presence of colchicine, colcemide, or vinblastine

Regeneration after 60 minRegeneration

Inhibitor mM Time added after 10 mm: Avg Standard Cellsavg flagellar flagellar .i. with no %length (jim) length devJAM) flagella Viable

(jim) (±Mm

None 3.9 11.6 1.6 0 74Colchicine 0.5 After shearing 3.0 11.2 2.7 4Colchicine 1.5 After shearing 3.0 4.0 2.1 10 81Colchicine 1.5 Before shearing 0.5 0.6 1.3 76 100Colchicine 6.0 After 10 min re- 3.3 8.1 3.7 10

generationColcemide 0.1 Before shearing 2.4 11.7 1.7 0Colcemide 0.3 Before shearing 1.0 11.6 1.8 2Colcemide 0.7 Before shearing 0 0.43 0.6 64Colcemide 0.7 After shearing 0.5 0.96 0.8 26Vinblastine 0.02 Before shearing 3.8 11.0 3.4 2Vinblastine 0.04 Before shearing 2.4 6.9 4.4 18Vinblastine 0.07 Before shearing 1.7 2.7 64Vinblastine 0.10 Before shearing 0 0 100 90Vinblastine 0.10 After 10 min re- 3.6 4.0 3.1 26

generation

62 J. BACTERIOL.

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 118,1974 CHLAMYDOMONAS MUTANTS RESISTANT TO ANTITUBULINS

TABLE 2. Effect of deoxycholate on coichicine inhibition of flagellar regeneration

RegenerationRegeneration after 60 min

Colchicinea Sodiume e after 10m: Standard Percent of(mM) deoxycholatea avg flagellar Avg flagellar deviation cells with %(mM)(%)

length (gm) length (pm) (±pim) no flagella Viable

0 0 3.6 11.6 1.6 0 743 0 0 0 1001.5 0 3.1 4.9 2.3 8 810.5 0 3.2 12.0 3.4 6 1000 0.04 1.3 10.0 3.5 6 840.5 0.04 0 0 1000.05 0.04 0 5.9 4.4 340.025 0.04 0 9.1 5.2 24

aAdded immediately after shearing.

TABLE 3. Millimolar concentrations of substances which inhibit growth or motility, or induce flagellardetachment, or inhibit regeneration after amputation

IInhibit growth Inhibit flagellar Cause flagellar InhibitSubstance regenerationa detachment motility

Yes ± No Yes _ No Yes ± No Yes 4 No

Colchicine 4.0 3.0 2.0 3.0 1.5 0.5 > 20.0 s 20.0Colcemide 1.0 0.3 0.1 0.7 0.4 0.1 5 5.0 > 5.0N-Desacetylthiocolchicine 1.0 0.4 0.2 1.0 0.1Desacetylcolchicine 2.0 0.6 1.0 0.2 0.06 1.0 0.3 0.08Colchicosamide 5 20.02-Methoxy-5-nitrotropone 0.017 0.005 0.002 > O0.1 b2-Hydroxy-5-nitrotropone 0.06 0.03 0.015Vinblastine 0.2 0.1 0.05 0.1 0.05 0.02 0.2 0.06 0.015EPT 0.6 0.4 0.3 0.04cSuccinylpodophyllotoxin (iso- 20.0 8.0 4.0 10.0 5.0 30.0d 10.0 5.0propylammonium salt)

Isopropylammonium succi- 8.0 4.0 2.0 50.0 20.0 20.0 10.0 1.0nate

Isopropyl N-phenyl carba- 0.3 0.06 0.02 > 1.0 0.1mate

Chloral hydrate 7.0 2.0 1.0 30.0 15.0 2.0 30.0 20.0 3.0Caffeine 4.0 2.0 1.0 s8.0 4.0Ni2+ acetate 0.15 0.10 0.05 0.3 0.10 0.03 > 0.6 0.18' 0.06

a Inhibitors were added to the cell suspension immediately before flagellar amputation.bAt this concentration flagella regenerated for 10 min, but by 60 min were again detached as the cells died and lysed.c Flagella became detached but the effect was transient and after 2 h they had regenerated.dPeripheral breakage after 60 min only.' Paralysis occurs only after 1 to 2 h.

potent in inhibiting growth. However, if a druginhibits growth and flagellar regeneration at thesame concentration, the result is also consistentwith these effects being mediated by inhibitionof tubulin assembly. Motility measurementswere incidental, since in this work we have notsystematically examined substances likely to bespecific paralytic agents. When we found thatsome drugs which seemed to inhibit flagellarregeneration were actually causing flagellar de-tachment under the same conditions, it wasclear that, even if they were established inhibi-tors of microtubule assembly, their toxic effectsin Chlamydomonas could not be attributed,

from our results, to this property. Flagellardetachment is a facile response to unfavorableenvironments. For example, cell suspensions inM lost all flagella when kept for 1 h at +4 C,though there was no loss of viability. At othertimes low temperature has produced popula-tions with flagella of unequal lengths; however,with this possible exception, disappearance offlagella (Table 3) seems always to have beendue to detachment rather than retraction (15).

Since growth inhibition was scored after 4 or 7days and the other parameters of Table 3 weremeasured during 1 h, slow permeation or de-composition of drugs might have influenced the

63

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

FLAVIN AND SLAUGHTER

respective results. Colchicine is rapidly decom-posed by ultraviolet light, but we found nodecomposition, as measured by the decrease inabsorbance at 352 nm, when a solution inmedium M was kept under growth lights for 2weeks. The "+" column under growth inhibi-tion usually indicates that some cell divisionoccurred followed by bleaching and death be-fore 7 days. Growth inhibition by podophyl-lotoxin or griseofulvin could not be studiedbecause of their relative insolubility in water.The "+" columns under other headings inTable 3 indicate drug concentrations whichallowed flagella to regenerate but to less than 11,um length in 60 min, or caused detachmentfrom, or paralysis of, some but not all cells.

Consistent with their ability to bind to tubu-lin and prevent assembly (13, 25, 27), colchicineand colcemide were equally effective in inhibit-ing growth or flagellar regeneration, and did notcause detachment or paralysis (Table 3). Col-chicosamide is inert, and likewise does not bindto tubulins (27). Desacetylcolchicine did notbind to tubulin in our hands (see Discussion),and its toxicity to Chlamydomonas may beunrelated to microtubules since low concentra-tions caused flagellar detachment. Methoxyni-trotropone inhibited growth only. Other drugswhich were primarily growth inhibitors, andwhose toxicity may therefore be unrelated totubulin assembly, were caffeine, isopropylN-phenyl carbamate, and possibly chloral hy-drate. Vinblastine binds strongly to tubulin (14)but its toxicity to Chlamydomonas may beunrelated, since low concentrations caused fla-gellar detachment.Podophyllotoxin also binds strongly to tubulin

(see Discussion). At 0.5 mM, the concentrationof a saturated aqueous solution, it caused littleor no inhibition of growth and flagellar regener-ation. The only water-soluble derivative ini-tially available to us, EPT, inhibited growth at0.6 mM, but caused a transient flagellar detach-ment at l/o this concentration (Table 3).However, at the latter low concentration podo-phyllotoxin itself was soluble and had no effect.Therefore the transient deflagellation caused by0.06 mM EPT was not due to its podophyl-lotoxin moiety, but may have been caused bythe thiourea moiety or an impurity, which wasdetoxified in the cells. We then prepared asecond water soluble derivative, succinylpodo-phyllotoxin. This inhibited growth, flagellarregeneration, and motility, but only at very highconcentrations. Moreover, the added moiety,isopropylammonium succinate, was itself moreinhibitory (Table 3). Twenty millimolar succinyl-

podophyllotoxin is 1.2% by weight and might beexpected to inhibit nonspecifically; only 0.2%acetate completely transforms the cells metabo-lism. Succinylation had apparently completelyabolished the affinity for tubulin, and this hasbeen confirmed by direct binding experiments(see Discussion).Though unrelated to the rest of this work, we

include in Table 3 some results with a nickelsalt, which intrigued us because nickel inhibitsmotility but activates the flagellar energy trans-ducing ATPase (5). Growth and flagellar regen-eration were also inhibited by the same concen-tration as motility. In one other experimentrelating to bivalent metals we found that [ethyl-enebis( oxyethylenenitrilo) ]tetraacetic acid(EGTA) had no effect on motility or flagellar re-generation of cells suspended in distilled water,but inhibited both if cells were in medium M(0.027 mM calcium).Conditions for mutagenesis and gene

expression. Since all of the antitubulins andrelated substances tested inhibited growth atleast as effectively as flagellar regeneration, itwas simpler to seek drug-resistant mutants onthe basis of ability to grow in the presence ofdrug concentrations which inhibited the wildtype, rather than on the basis of ability toregenerate flagella and swim in the presence ofinhibitors. In preliminary experiments to opti-mize mutagenesis conditions we scored for abil-ity to grow in the presence of streptomycin (9).The ratio of mutation frequency to lethality washigher with methyl than ethyl methanesulfo-nate, as previously reported for arg-1 reversionin Chlamydomonas (11), and was comparable tothat obtained with N-methyl-N'-nitro-N-nitrosoguanidine (Table 4). Since the latterselectively mutates replicating fornis of deoxyri-bonucleic acid (9) and may produce mutationswhich later segregate anomalously, we chosemethyl methanesulfonate for further mutagene-SiS.The recovery of induced strr (streptomycin

resistant) mutants has been shown to be in-creased if the cells are kept for 15 to 20 h undernonselective conditions after mutagenesis (9).The results of Table 5 confirm this conclusionand also show that the strr mutants were notdepleted after several generations of growth inminimal medium, but were evidently able tocompete effectively with wild type. Transfer ofindividual colonies, with sterile toothpicks, toagar plates with and without streptomycinconfirmed that at least 75% were mutants.Table 5 shows some very tentative results on thefrequency of mutation to colchicine and vin-

64A J . BACTERIOL

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 118,1974 CHLAMYDOMONAS MUTANTS RESISTANT TO ANTITUBULINS

TABLE 4. Mutation frequency and viability undervarious mutagenizing conditions

Strepto-mycin-

Concn or % resistantMutagenizing conditions time of Viable mutants

exposure per 101viablecells

N-Methyl-N'-nitro-N- 0 gg/ml 70.0 5.0nitrosoguanidine, 5 Ag/ml 54.0 15.030 min 15 Ag/ml 51.0 25.0

50 Ag/ml 13.0 46.0Ethyl methanesulfo- 30 min 37.0 2.5

nate, 0.23 M 60 min 13.0 5.0120 min 0.06 < 120.0

Methyl methanesulfo- 0 M 84.0 0.5nate, 120 min 0.013M 39.0 40.0

0.015M 11.0 41.00.017M 0.9 -60.00.020M 0.03

blastine resistance. Colchr (colchicine resistant)mutants were comparable to strr in frequency,and also were not overgrown by wild type duringseveral generations in minimal medium. Vblr(vinblastine resistant) mutants appeared to beincreased after a period of gene expressionwithout growth, but to be overgrown by wildtype under growth conditions. The figures forvinblastine resistance are not very significant,since of the four "mutants" later examined indetail two were indistinguishable from wild type(see below).Mutants resistant to antitubulins. Mu-

tants chosen for further study originated fromthree different experiments and after variousconditions of gene expression. The scoring ofmutants on petri plates containing colchicineand related drugs was not as straightforward asin the case of streptomycin. On plates contain-ing 10 mM colchicine wild-type cells underwentresidual growth, did not bleach promptly, andsometimes escaped inhibition altogether. Thecolchr mutant colonies invariably died between

10 and 18 days on the original plates and had tobe picked under the dissecting microscope witha block of agar, when they were quite small.Some of these difficulties may have been due tousing drug concentrations barely sufficient toinhibit a very heavy inoculum of wild-typecells, which was dictated primarily by theexpense or limited supply of the drugs.Three colchr and four vblr colonies were

picked, cloned as described in Materials andMethods, and examined for their ability to growand regenerate flagella in the presence of colchi-cine or vinblastine, or grow and remain motilein the presence of EPT (Tables 6 to 8). Theresults for two vblr mutants were indistinguish-able from those for wild-type controls and havenot been included. Three EPTr (EPT-resistant)colonies were picked but were not cloned be-cause of the very small amount of EPT availa-ble; after one subculture in minimal liquidmedium they all appeared slightly resistant toEPT inhibition of growth when tested as inTable 8. After further subculture, however, theyall scored as wild types. It was realized that thewild-type background cells had remained viableon the original EPT plate, although they did notdivide, because they could be seen to be motilewith the dissecting microscope, and that thesehad probably contaminated and later over-

grown the mutant cultures. The original resultswith one of these mutants were sufficientlyintriguing to be included in the tables despitethis uncertainty: EPTr-2 scored as cross-resist-ant to vinblastine, but hypersensitive to colchi-cine.No mutants were obtained when a total of

about 106 viable cells were plated on mediumcontaining 0.8 mM isopropyl N-phenyl carba-mate or 0.5 mM nickel acetate.Of the three colchicine-resistant mutants,

colchr_1 grew in the presence of 10 times highercolchicine concentrations than were toleratedby wild type, and could also regenerate flagellain high concentrations of colchicine (Table 6). Itwas strikingly cross-resistant to EPT inhibition

TABLE 5. Mutation frequency after various periods of gene expression

Mutants, per 10' viable cells, resistant to:

Expt conditions % Viable Streptomycin Colchicine Vinblastine(50 jg/ml) (10 mM) (0.5 mM)

Before mutagenesis 66 2 0.4 < 1Treatment with 0.016M MMS for 2 h 34 13 2As in (2) followed by 20 h incubation with- 64 18 60

out cell increaseAs in (2) followed by 48 h incubation with 100 26 28.0 < 1

20-fold cell increase

65

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

FLAVIN AND SLAUGHTER

TABLE 6. Drug-resistant mutants: effect of colchicine on growth and flagellar regeneration

Flagellar regenerationmM concn inhibiting

Strain growth Avg flagellar length, in microns, after 60 min regeneration inpresence of millimolar concna:

Yes No 0 0.8 1.5 2.0 3.0 3.5 4.0 4.5

Wild type 4 3 2 11 4.8 0.1 0Colchr-1 40 30, 20 10 8.5 8.8 8.6 8.6 4.9 2.0 1.2Colchr-2 10 8 6 7.9 3.8 1.2Colchr-3 10 5 11 7.4 6.2 3.4Vbl r-I 4 3 2Vblr-4 4 3 2 12 6.0 1.7 0EPTr -2s

a Colchicine was added immediately before flagellar amputation.This mutant was subsequently lost.

TABLE 7. Drug-resistant mutants: effect ofvinblastine on growth and flagellar regeneration

Flagellar regeneration

mM concn inhib- Avg flagellar length,Strain iting growth in microns, after 60 min

regeneration in presenceof millimolar concna:

Yes No 0 0.03 0.04 0.05

Wild type 0.2 0.1 0.05 1 1 4.1 1.0 0Colchr-1 0.3 0.2 0.1 11.0 10.0 0Colchr-2 0.2 0.1 0.05 7.3 5.0 0Colchr-3 0.3 0.2 0.1 4.7 1.8Vblr-1 0.3 0.2 0.1 6.8 0Vblr-4 0.2 0.1 0.05 2.1 0.1EPTr-2 0.3 0.15

a Vinblastine was added immediately before flagel-lar amputation.

of growth (Table 8), but not to the transientdeflagellation caused by very low concentra-tions (all the latter results in Table 8 areprobably within the wild-type range exceptthose for vblr-4). This is consistent with theconclusion that growth inhibition by EPT iscaused by the podophyllotoxin moiety but de-flagellation is due to a labile impurity or to thethiourea moiety. Colchr_1 was also slightlycross-resistant to inhibition of growth and fla-gellar regeneration by vinblastine (Table 7).Colchr-2 and colchr-3 were resistant to onlymoderate concentrations of colchicine, andshowed equivocal or no resistance to the otherdrugs.Two of the four vblr mutants were indistin-

guishable from wild type and the other twoscored equivocally, vbl4-1 as slightly resistant tovinblastine and vblr-4 as slightly resistant toEPT.

Colchr_1 grew somewhat slowly in minimalmedium and the liberation of daughter cellsfrom their maternal envelope was delayed.

DISCUSSIONEffects of antitubulins and related sub-

stances on Chlamydomonas. Colchicine wasequally effective as inhibitor of growth andflagellar regeneration. As these processes bothdepend on generation of new microtubules, andas colchicine showed no indication of affectingprocesses (motility, flagellar detachment) un-related to microtubule assembly, the results ofTable 3 are consistent with the conclusion thatits effects on Chlamydomonas were due toantitubulin activity rather than some otherunknown toxicity. Colchicine binds to cytoplas-mic tubulin from mammalian brain (1 mol per110,000 molecular weight dimer) with an appar-ent dissociation constant of about 2 x 10-6 M(24, 27), but 1,000 times higher concentrationwas required to inhibit Chlamydomonas. It hasbeen difficult to show binding to flagellar tubu-lin and the latter might have a lower affinity(25). It seems more likely that colchicine is not

TABLE 8. Drug-resistant mutants: effect of EPT ongrowth and motility

mMconcninhib % Non-motileconihi cells after 60 min

Strain in presence of:

Yes No 0.04 0.08mm mm

Wild type 0.4 0.3 0.2 94 100Colchr_1 1.5 0.8 0.5 63Colchr-2 0.4 0.3 0.2 100Colchr-3 0.4 0.2 74 100Vblr-4 0.6 0.3 0 50EPTr-2 0.5 0.3

66 J. BACTERIOL.

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 118,1974 CHLAMYDOMONAS MUTANTS RESISTANT TO ANTITUBULINS

easily permeable, as suggested by the observa-tion that detergents can increase its effective-ness (Table 2). To confirm whether the an-titubulin activity of various substances was

indeed responsible for their effects on

Chiamydomonas, we have compared the latterwith their ability to inhibit binding of tritium-labeled colchicine to brain tubulin, and some

preliminary results, to be described in detailelsewhere, are outlined in Table 9. Bindingexperiments were done with tubulin purifiedfrom pig brain by the batch procedure (22), andbound ligand was determined by modificationsof the DE-81 paper filtration method (22).We also studied the ability of drug derivatives

to inhibit the binding of 3H-podophyllotoxin(Table 9), prepared from 4'-demethylpodophyl-lotoxin (6). Previous studies of the inhibition of3H-colchicine binding by unlabeled podophyl-lotoxin (23) suggested that they bind to thesame site. The results of Table 9 (top) confirmthat both ligands compete for the same site.Saturation with ligand indicated the samenumber of sites, although there was probablyalso some nonspecific binding of podophyl-

lotoxin at high concentrations. The dissociationconstant for podophyllotoxin was calculated tobe about 7 x 1O- M, from Scatchard plots; theapparent rates of association and dissociationwere about equal at 37 C, and much faster thanthose of colchicine. Unexpectedly, the compara-ble rates for 9H-colcemide were similar to thosefor podophyllotoxin; total sites available tocolcemide were the same as for colchicine,contrary to our preliminary report (Fed. Proc.32:642, 1973).Comparison of the results of Tables 3 and 9

for various colchicine derivatives confirms thatonly those which specifically inhibit growth andflagellar regeneration (colchicine, colcemide)bind to brain tubulin. Colchicosamide was inac-tive in both cases. Desacetylcolchicine causedflagellar detachment suggesting a possible tox-icity unrelated to antitubulin activity, and thesame sample did not effectively inhibit binding.Others have reported that this derivative doesinhibit colchicine binding (27); in any case theresults with the sample available to us were

consistent. Methoxynitrotropone, a possible an-

alogue of the tropolone C ring of colchicine,

TABLE 9. Inhibition of binding of labeled colchicine or podophyllotoxin to brain tubulin, by various derivativesand analogues

% Inhibition of binding of:

Inhibitor concn Prior incubationInhibitora ~~~(M x 10-') with inhibitor3'H-Colchicine 3Hlotophin(min) (6 x 10-'M) (2 x iOnM)

Colchicine 12 0 74Colchicine 12 60 95Podophyllotoxin 12 0 83Podophyllotoxin 12 60 85Colcemide 12 0 47 22Colcemide 12 60 45 39Colcemide 120 0 78 68Desacetylcolchicine 18 0 0 1Desacetylcolchicine 60 0 23 16Colchicosamide 12 60 9 0Colchicosamide 30 60 4 0Methoxynitrotropone 30 0 0 5EPT 12 0 34EPT 20 0 46 49EPT 30 0 61 61Epipodophyllotoxin 18 0 22 26Epipodophyllotoxin 60 0 56 47Succinylpodophyllotoxin 12 0 4 0Succinylpodophyllotoxin 30 0 0 144'-Demethylpodophyllotoxin 12 0 77 674 -Demethylpodophyllotoxin 30 0 89 79Picropodophyllin 12 0 2 0Picropodophyllin 30 0 2 6

a The following had no significant effect, at concentrations of 120 x 10 M, on the binding of colchicine or

podophyllotoxin: vinblastine, chloral hydrate, caffeine, isopropyl N-phenyl carbamate, griseofulvin, andcytochalasin B.

67

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

FLAVIN AND SLAUGHTER

inhibited growth only, and did not inhibitbinding. Its methoxy group is activated by thenitro substituent and is readily displaced bytubulin amino groups (T. Watanabe and M.Flavin, Fed. Proc. 32:642, 1973); colchicinebinding cannot involve this reaction since theanalogous methoxy group is the location of theisotopic label.

Similarly, with the water soluble podophyl-lotoxin derivatives, EPT inhibited binding(Table 9) and also growth; flagellar regenerationcould not be tested because something otherthan the podophyllotoxin moiety caused de-flagellation. Epipodophyllotoxin itself also in-hibited binding, possibly less well than EPT.EPT was an unsatisfactory derivative becauseof the transient toxicity to Chlamydomonasunrelated to antitubulin activity. We thereforesynthesized a second water-soluble derivative,succinylpodophyllotoxin. This proved to beinactive towards both Chlamydomonas and tu-bulin binding. Inactivation by such a minorstructural modification was disappointing toour future interest in affinity labeling withpodophyllotoxin derivatives, but it is not sur-

prising that a changed group should preventbinding to a presumably hydrophobic, protein-protein interaction site. Succinylpodophyl-lotoxin may be useful, in evaluating in vivo drugeffects, as a water-soluble derivative with no

antitubulin activity. Picropodophyllin (6) issuitable for the same purpose (Table 9), where a

low water solubility is sufficient.None of the remaining drugs affected

Chlamydomonas in a manner expected for an-

titubulins (Table 3), and none inhibited colchi-cine or podophyllotoxin binding (Table 9). Thisdoes not entirely preclude binding at a separatesite, as is the case with vinblastine. Vinblastineis an example of an antitubulin whose toxicityto Chlamydomonas is probably unrelated tomicrotubules, because some other process ismore sensitive to inhibition. Isopropyl N-phenylcarbamate inhibited growth much more thanflagellar regeneration (Table 3). The herbicideallows microtubules to assemble in plant cellsbut in a disorganized orientation (7). Sinceshearing does not remove the basal bodies at theflagellar origin, one might expect that flagellarregeneration would be insensitive. Chloral hy-drate and caffeine were also primarily growthinhibitors. Chloral hydrate has been classifiedas a mitotic spindle poison (page 118 in refer-ence 3). Caffeine had been reported to inhibitflagellar regeneration and a Chlamydomonascyclic AMP phosphodiesterase (16). The resultson inhibition of flagellar regeneration and mo-

tility by Ni2", or Ca2" plus EGTA, pertain moreto a separate interest in motility and its regula-tion.Chlamydomonas mutants resistant to an-

titubulin drugs. The distribution of micro-tubules in Chlamydomonas suggests that thetargets of antitubulin drugs should be mitosis,cell and nuclear cleavage, and the generation offlagella. This work bears on the feasibility ofobtaining drug resistant mutants. Such mu-tants could have decreased permeability to, orincreased detoxification of, the drugs, or achange in the quantity or quality of tubulin,specifically an amino acid substitution allowingmicrotubule assembly but not antitubulin bind-ing. The latter type of mutant would seem lesslikely to be found if antitubulin binding capac-ity were shown to have been conserved inevolution, but this is not established.No mutants were obtained resistant to isopro-

pyl N-phenyl carbamate, which is not strictlyan antitubulin, but further trials are warrantedas we examined only 106 viable cells. Vinblas-tine is an authenic antitubulin but its adverseeffects on Chlamydomonas are probably due tosome other toxicity. Mutation to drug resistanceseemed to be frequent (Table 5), but of the fourmutants that were cloned two proved to beindistinguishable from wild type and twoshowed very low resistance. Moreover, vblr-4was resistant to EPT, not vinblastine, and itwas resistant to the transient deflagellationeffect which is not caused by the podophyl-lotoxin moiety. Three presumptive mutantspicked from EPT plates, but not cloned, werefound at first to have a slight resistance toinhibition of growth by EPT, but this propertywas later lost, perhaps because of contamina-tion with wild type. EPTr-2 scored as slightlycross-resistant to vinblastine but hypersensitiveto colchicine (Table 6), a phenotype which isquite suggestive of an altered primary sequencein tubulin. This study is therefore worth pursu-ing, but preferably with a different water-solu-ble derivative of podophyllotoxin.

Colcemide-resistant mutants of a fissionyeast have been previously reported; colcemide,but not colchicine, inhibited cell cleavage inthis organism (18). Colchicine-resistant mu-tants of Chlamydomonas have also been iso-lated by D. J. L. Luck (personal communica-tion) and by Adams and Warr (1) who found theresistance to be Mendelian, and to segregatewith a property of slow growth in basal medium.We found that methyl methanesulfonate in-duced a high incidence of mutation to colchi-cine resistance (Table 5), and all three mutants

J. BACTERIOL.68

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 118, 1974 CHLAMYDOMONAS MUTANTS RESISTANT TO ANTITUBULINS

cloned proved to be authentic. These mutantshave been examined so far only for their abilityto grow and regenerate flagella in the presenceof each of three antitubulin drugs (Tables 6 to8). Colchr-2 and colchr-3 were moderately resist-ant to colchicine in both respects, and had littleor no resistance to EPT or vinblastine. Colchr_1was resistant to high concentrations of colchi-cine and was also strikingly cross-resistant tothe growth inhibiting effect of EPT. Mutants ofthis type would also be worth examining forpossible alterations in tubulin.

ACKNOWLEDGMENTSWe are grateful, for helping to introduce us to unfamiliar

organelles and organisms, to M. Shelanski, J. L. Rosenbaum,R. P. Levine, and N. Gillham.

LITERATURE CITED

1. Adams, M., and J. R. Warr. 1972. Colchicine-resistantmutants of Chiamydomonas reinhardtii. Exp. Cell Res.71:473-475.

2. Blum, J. J., and J. Lubliner. 1973. Biophysics of flagellarmotility. Annu. Rev. Biophys. Bioeng. 2:181-218.

3. Deysson, G. 1969. Antimitotic substances. Int. Rev.Cytol. 24:99-148.

4. Fell, K. R., and D. Ramsden. 1967. Colchicum: a reviewof colchicums and the sources, chemistry, biogenesisand assay of colchicine and its congeners. Lloydia30:123-140.

5. Gibbons, B. H., and I. R. Gibbons. 1972. Flagellarmovement and adenosine triphosphatase activity insea urchin sperm extracted with Triton X-100. J. CellBiol. 54:75-97.

6. Hartwell, J. L., and A. W. Schrecker. 1958. The chemis-try of podophyllum. Fortschr. Chem. Org. Naturst.15:83-166.

7. Hepler, P. K., and W. T. Jackson. 1969. IsopropylN-phenyl carbamate affects spindle microtubule orien-tation in dividing endosperm cells of HaemanthusKatherinae Baker. J. Cell Sci. 5:727-743.

8. Johnson, U. G., and K. R. Porter. 1968. Fine structure ofcell division in Chiamydomonas reinhardtii. J. CellBiol. 38:403-425.

9. Lee, R. W., and R. F. Jones. 1973. Induction of Mendel-lian and non-Mendellian streptomycin resistant mu-tants during the synchronous cell cycle ofChiamydomonas reinhardtii. Mol. Gen. Genet.121:99-108.

10. Levine, R. P. 1971. Preparation and properties of mutantstrains of Chlamydomonas reinhardtii, p. 119-129. InA. San Pietro (ed.), Methods in enzymology, vol. 23.Academic Press Inc., New York.

11. Loppes, R. 1969. Effect of the selective medium on themanifestation of mutations induced with mono-

alkylating agents in Chiamydomonas reinhardtii. Mu-tat. Res. 7:25-34.

12. McVittie, A. 1972. Genetic studies on flagellar mutants ofChiamydomonas reinhardtii. Genet. Res. 9:157-164.

13. Olmsted, J. B., and G. G. Borisy. 1973. Microtubules.Annu. Rev. Biochem. 42:507-540.

14. Owellen, R. J., A. H. Owens, Jr., and D. W. Donigan.1972. The binding of vincristine, vinblastine and col-chicine to tubulin. Biochem. Biophys. Res. Commun.47:685-691.

15. Rosenbaum, J. L., J. E. Moulder, and D. L. Ringo. 1969.Flagellar elongation and shortening in Chiamydomonas.J. Cell Biol. 41:600-619.

16. Rubin, R. W., and P. Filmer. 1973. Adenosine 3', 5'-cyclicmonophosphate in Chiamydomonas reinhardtii. J. CellBiol. 56:628-635.

17. Schrecker, A. W., P. B. Maury, J. L. Hartwell, and J.Leiter. 1955. Components of podophyllin: ionic deriva-tives of podophyllotoxin and of the peltatins. J. Amer.Chem. Soc. 77:6565-6568.

18. Stetten, G., and S. Lederberg. 1970. Colcemide sensitiv-ity of fission yeast and the isolation of colcemide-resist-ant mutants. Science 168:485.

19. Sueoka, N., K. S. Chiang, and J. R. Kates. 1967.Deoxyribonucleic acid replication in meiosis ofChlamydomonas reinhardtii: isotope transfer experi-ments with a strain producing eight zoospores. J. Mol.Biol. 25:47-66.

20. Warner, F. D. 1972. Macromolecular organization ofeukaryotic cilia and flagella. Advan. Cell Mol. Biol.2:193-235.

21. Watanabe, T., and M. Flavin. 1973. Two types ofadenosine triphosphatase from flagella ofChlamydomonas reinhardtii. Biochem. Biophys. Res.Commun. 52:195-201.

22. Weisenberg, R. C., G. G. Borisy, and E. W. Taylor. 1968.The colchicine binding protein of mammalian brainand its relation to microtubules. Biochemistry7:4466-4479.

23. Wilson, L. 1970. Properties of colchicine binding proteinfrom chick embryo brain: interactions with vincaalkaloids and podophyllotoxin. Biochemistry9:4999-5007.

24. Wilson, L., and J. Bryan. 1974. Biochemical and pharma-cological properties of microtubules. Advan. Cell Mol.Biol., in press.

25. Wilson, L., and I. Meza. 1973. The mechanism of actionof colchicine: colchicine binding properties of seaurchin sperm tail outer doublet tubulin. J. Cell Biol.58:709-719.

26. Witman, G. B., K. Carlson, J. Berliner, and J. L.Rosenbaum. 1972. Chlamydomonas flagella: isolationand electrophoretic analysis of microtubules, matrix,membranes, and mastigonemes. J. Cell Biol.54:507-539.

27. Zweig, M. H., and C. F. Chignell. 1973. Interaction ofsome colchicine analogues, vinblastine and podophyl-lotoxin with rat brain microtubule protein. Biochem.Pharmacol. 22:2141-2150.

69

on February 16, 2020 by guest

http://jb.asm.org/

Dow

nloaded from